{"paper":{"title":"Neural Decision-Propagation for Answer Set Programming","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Decision-propagation computes stable models by alternating falsity decisions and truth propagations, and its neural version learns to do so efficiently.","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Katsumi Inoue, Sota Moriyama, Thomas Eiter","submitted_at":"2026-05-03T09:22:26Z","abstract_excerpt":"Integration of Answer Set Programming (ASP) with neural networks has emerged as a promising tool in Neuro-symbolic AI. While existing approaches extend the capabilities of ASP to real world domains, their reasoning pipelines depend on classical solvers, which is a bottleneck for scalability. To tackle this problem, we propose a new method to compute stable models, called decision-propagation (DProp), which alternates falsity decisions and truth propagations. Successful DProp computations are shown to capture the stable model semantics. We then develop Neural DProp (NDProp), a differentiable ex"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Successful DProp computations are shown to capture the stable model semantics. NDProp can learn to efficiently compute stable models, and it improves accuracy and scalability on neuro-symbolic benchmarks.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That neural decisions combined with fuzzy propagations can reliably approximate exact stable-model computation while preserving correctness and generalizing beyond the training distribution used in the benchmarks.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"NDProp learns decision heuristics via neural networks and fuzzy propagation to compute stable models in ASP, improving accuracy and scalability over prior neuro-symbolic methods.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Decision-propagation computes stable models by alternating falsity decisions and truth propagations, and its neural version learns to do so efficiently.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"1ea4d7e85b17a7e5c005bddbafeff24a1d03251db439d9d604a4a5604b31c7bc"},"source":{"id":"2605.01797","kind":"arxiv","version":2},"verdict":{"id":"aae59194-e594-4753-904f-9b56522f8db6","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-10T15:08:45.540414Z","strongest_claim":"Successful DProp computations are shown to capture the stable model semantics. NDProp can learn to efficiently compute stable models, and it improves accuracy and scalability on neuro-symbolic benchmarks.","one_line_summary":"NDProp learns decision heuristics via neural networks and fuzzy propagation to compute stable models in ASP, improving accuracy and scalability over prior neuro-symbolic methods.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That neural decisions combined with fuzzy propagations can reliably approximate exact stable-model computation while preserving correctness and generalizing beyond the training distribution used in the benchmarks.","pith_extraction_headline":"Decision-propagation computes stable models by alternating falsity decisions and truth propagations, and its neural version learns to do so efficiently."},"integrity":{"clean":false,"summary":{"advisory":2,"critical":0,"by_detector":{"doi_compliance":{"total":2,"advisory":2,"critical":0,"informational":0}},"informational":0},"endpoint":"/pith/2605.01797/integrity.json","findings":[{"note":"DOI in the printed bibliography is fragmented by whitespace or line breaks. A longer candidate (10.1007/978-3-540-24599-5\\_17) was visible in the surrounding text but could not be confirmed against doi.org as printed.","detector":"doi_compliance","severity":"advisory","ref_index":36,"audited_at":"2026-05-19T16:58:34.590334Z","detected_doi":"10.1007/978-3-540-24599-5\\_17","finding_type":"recoverable_identifier","verdict_class":"incontrovertible","detected_arxiv_id":null},{"note":"DOI in the printed bibliography is fragmented by whitespace or line breaks. A longer candidate (10.1007/978-3-540-28633-2\\_7) was visible in the surrounding text but could not be confirmed against doi.org as printed.","detector":"doi_compliance","severity":"advisory","ref_index":17,"audited_at":"2026-05-19T16:58:34.590334Z","detected_doi":"10.1007/978-3-540-28633-2\\_7","finding_type":"recoverable_identifier","verdict_class":"incontrovertible","detected_arxiv_id":null}],"available":true,"detectors_run":[{"name":"ai_meta_artifact","ran_at":"2026-05-20T17:35:56.431865Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_title_agreement","ran_at":"2026-05-20T05:01:22.689447Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T16:58:34.590334Z","status":"completed","version":"1.0.0","findings_count":2}],"snapshot_sha256":"61f8e2bb53afe06f8875f711e1c9da804da550d0fdc6023697edf3cb4d675534"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}